Thread: A Trillion frames per second camera.

Originally Posted by tommo

How did you troll me exactly?

Anyway, for once I can say the Americans have a better system. That Continental European system is just fucking retarded.

Technically the modern system is just a dumbed down version of the older system. You only think it's stupid, because you're used to the modern system.

This kind of reporting just makes me hate news companies. I read this article on the BBC and their headline was 'trillion frame per second camera can trick light!!!!' or something, and then it took almost half the article before they mentioned that actually, it doesn't record a trillion frames per second, it amalgamates many many shoots and then plays a trillion frames per second.

This is why you need actual scientists reporting on science stories.

Originally Posted by Marvo

Technically the modern system is just a dumbed down version of the older system. You only think it's stupid, because you're used to the modern system.

Probably. But Million, Trillion, Billion etc. just seems to make more logical sense, because bi = two, tri = 3 etc. and you can easily deduce what the next one will be called.

1000^3 = bi
1000^4 = tri
1000^5 = quadri

Sure is logical in here.

Originally Posted by tommo

Probably. But Million, Trillion, Billion etc. just seems to make more logical sense, because bi = two, tri = 3 etc. and you can easily deduce what the next one will be called.

You can work out the next name, sure, but the names have absolutely nothing to do with the values they describe. The system works just as well in the old version as the new version. The advantage to the old system, is that you only need to know half the names, in order to decsribe the same amount of large number.

For example, in order to describe 4 kinds of large numbers in American English, you need to know the words "Million, billion, trillion, quadrillion", in the old system, you only need to know "Million, billion", from there you can work the rest out with ease (milliard, billiard).

Originally Posted by Xei

1000^3 = bi
1000^4 = tri
1000^5 = quadri

Sure is logical in here.

Well it isn't if you think of it like that. But you know that, that's why you purposely did it that way.

Uh... how else can you think about it? That's how that naming scheme works, in powers of 1000. You only think it's particularly logical out of familiarity.

Originally Posted by Xei

Uh... how else can you think about it? That's how that naming scheme works, in powers of 1000. You only think it's particularly logical out of familiarity.

1000 x 1000 = million
1000 x 1m = billion
1000 x 1b = trillion
etc.

I was about to post I already made a thread about it. Then saw that a another another person had beat me to it. Damn.

Well I find this extremely amazing. Like, mind blowlingly awesome batshit holy fuck. I spammed my physics teacher with links to youtube and the article and he didn't seem to give half the shits I gave.

This video just has so much cool shit in it, the most fascinating bit I think is the one at 2:00, where light is being shot through a container with water and into grating, which then splits the light as it reflects back.

Originally Posted by tommo

1000 x 1000 = million
1000 x 1m = billion
1000 x 1b = trillion
etc.

1,000,000 = million
1,000,000*1m = billion
1,000,000*1b = trillion

ROFL how absolutely absurd!

Originally Posted by Darkmatters

So yeah - you can record at the equivalent of a trillion FPS - as long as you can precisely repeat the event over and over again. Not much of a trick if you think about it.

I'd say it's a pretty neat trick to be able to record fast enough for us to be able to see light moving, even if it is only a line. And it might lead us to better sensors at a later stage.

Originally Posted by Marvo

Nope.

Million = Million
Milliard = Billion
Billion = Trillion
Billiard = Quadrillion
Trillion = Quintillion
Trilliard = Sextillion

List of numbers - Wikipedia, the free encyclopedia

Why can't we just use the SI prefixes for everything? "Yes, I won a mega-dollar!"

Originally Posted by khh

I'd say it's a pretty neat trick to be able to record fast enough for us to be able to see light moving, even if it is only a line. And it might lead us to better sensors at a later stage.

If that in fact is what they're doing, then yes. But it sounds like they're essentially just creating the illusion of super-high speed photography. I'm not sure - like I said, I'd need to know more about how it works. But understand - what they're doing is creating a whole series of light flashes and using many of them combined together to create this imagery. It could be as simple as having a strobe light and a camera that's set to strobe slightly slower or slightly faster, so that one time you get the beginning of a flash, next time you get just a millisecond (or whatever - trillisecond?) later - of the NEXT flash, and so on. Then you can combine those images together into something that essentially mimiks what you'd get if you were actually able to film a single light flash at super speed.

I this is what they're doing, I can certainly see how it can help scientists study light - since each flash is presumably behaving exactly like all the others, so the information you get is almost the same as recording a single flash at super-speed.

However, if this IS how it works, then the technique isn't good for much else. Only events that are absolutely repeatable with no variation whatsoever from one to the next.

I should probably read up on it before I say these things - possibly they do describe the mechanism in better detail.

Ok, read up on it and it's what I thought. I guess to me this isn't anything new because it's what I do as an animator - I use a still camera and create the illusion of movement by reaching in between frames to move my puppets, and when I string the images together it looks like they're moving around all by themselves.

Though this is an innovative new way to use similar technology. But I think a lot of people are confused and think the camera is somehow actually recording things at immensely fast speeds.

As I said before, I think it is an incredible research tool. Just not as innovative or new as it sounds, and not useful for anything that isn't absolutely repeatable.

Originally Posted by Darkmatters

If that in fact is what they're doing, then yes. But it sounds like they're essentially just creating the illusion of super-high speed photography.

That was how I interpreted the explanation in the video in O's post, but it seems you've found you're right. Watching the video anew, I noticed he said at the beginning that they'd made a "virtual slow-motion camera". So yeah, I guess it's not as impressive as it sounded.

Originally Posted by Darkmatters

millisecond (or whatever - trillisecond?)

deci (d): 10^-1
centi (c) 10^-2
milli (m): 10^-3
micro (µ): 10^-6
nano (n): 10^-9
pico (p): 10^-12
femto (f): 10^-15
atto (a): 10^-18
zepto (z): 10^-21
yocto (y): 10^-24

I guess the it's nanosecond here.

I found the information on a BBC website about the camera: BBC News - MIT's trillion frames per second light-tracking camera

I'll paste in the relevant parts:

Direct recording of light is impossible at that speed, so the camera takes millions of repeated scans to recreate each image...

To create the technique, the scientists adapted a "streak tube" - equipment used to take data readings from light pulses.

It works in a similar fashion to the way pictures are created on traditional television cathode ray tubes, scanning one thin horizontal line at a time.

Since each image is only equivalent to one scan line on the television set, many hundred scans had to be taken to create a single frame.

The scientists did this by repeating each shot, angling the camera's view with mirrors to record a different scan line of the object.

As a result, the technique is only suitable for capturing an event that can be recreated exactly the same way multiple times.